Exploring factors influencing urban sprawl and land-use changes analysis using systematic points and random forest classification

This study examines urban sprawl and land-use changes by utilizing systematic points and random forest classification. The research focuses on Neyriz city in Fars Province, Iran, using satellite images from 1986 to 2016. Land-use maps were classified into urban, mountains, bare land, and vegetation using random forest machine learning in Google Earth Engine. Seven factors were analyzed in the geographic information system, and a kernel analysis of systematic points (KASyP) was applied to rank spatial variables. A grid of 4300 systematic points with 90 x 90 m spacing was created for data extraction and scatter plot generation. The study predicts a 12.3% increase in urban areas by 2026, with significant land changes near commercial, educational, administrative, and road areas. KASyP shows low change probability in mountains and high change probability in bare land. Notably, bare land to urban changes were prominent along roads and rivers. This research assists land use planners by identifying influential driver factors for land-use changes. It highlights the need to consider spatial variables and long-term trends in land-use analysis to mitigate risks, resolve conflicts, improve ecological safety, and maximize land potential. The combination of systematic points and random forest classification provides a robust methodology for managing urban sprawl and its environmental implications.